QSerialport readyRead() for measuring revolutions

[Ronel_qtmaster]

@herrgross now how is your code in arduino?

[herrgross]

@Ronel_qtmaster here's the complete Arduino code:

char incomingBytes[2];

int motorPin = 3;  // motor connected to digital pin 3 (PWM)
int bremsPin = 13;
// Drehzahlmessung
int interruptPin = 2;  // Hallsensor to pin 2
int pwmsend = 0;

void setup() {
  pinMode(motorPin, OUTPUT);
  //Löschen der Timer/Counter Control Register A und B
  TCCR1A = 0;
  TCCR1B = 0;
  //Modus Phase correct PWM-Mode 10 Bit einstellen
  TCCR1A |= (1 << WGM10) | (1 << WGM11);
  //Vorteiler auf 8 setzen
  TCCR1B |= (1 << CS11);
  //Nichtinvertiertes PWM-Signal setzen
  TCCR1A |= (1 << COM1A1); 
  //PWM-Pin 9 als Ausgang definieren
  DDRB |= (1 << DDB1);

  pinMode(bremsPin, OUTPUT);
  pinMode(interruptPin, INPUT_PULLUP);
  attachInterrupt(digitalPinToInterrupt(interruptPin), impuls, HIGH);

  Serial.begin(9600);
}


void loop() {
  OCR1A = 00;
  if (Serial.available()) {
    Serial.readBytes(incomingBytes, 2);
    if (incomingBytes[0] == 4) {
      //analogWrite(motorPin, 0);
      OCR1A = 0;           //setzt pin9 auf 0
      digitalWrite(bremsPin, HIGH);
    };
    // pwm senden
    if (incomingBytes[0] != 4) {
      int pwmsendnew;
      pwmsendnew = int(incomingBytes[0])*256 + int(incomingBytes[1]);
      if (pwmsendnew != pwmsend) {
        pwmsend = pwmsendnew;
        digitalWrite(bremsPin, LOW);
        analogWrite(motorPin, pwmsend);
        OCR1A = pwmsend;
      };
    };
  };
};

void impuls() {
    if (Serial.availableForWrite()) {
      Serial.write(1);
    };
};

[Ronel_qtmaster]

@herrgross reading your code remind me of something.You have that doubling behavior because you consider the values sent by arduino as byte and it does not take the entire values for intance the next line character "/r/n"

read all the serial values per line and as String and the doubling behavior will go

[herrgross]

@Ronel_qtmaster thank you for the nice idea ... but it did not solve the problem..... anyhow I tried to control what's actually arriving with readall() into a QByteArray and the checking the size of the array and reading the data. There was nothing arriving than the '1' sent by arduino, no endline character or other

[Ronel_qtmaster]

@herrgross okay.I was only pointing that the data sent to qt through serial port have to be a string to avoid doubling.Also , it is not recommend to read data from both arduino and qt.One should send and the other receive.

[JonB]

@herrgross said in QSerialport readyRead() for measuring revolutions:

Looking at this list I find that dataRx seems to be called twice per byte sent by the ARDUINO

Can you show a minimal example? For example, you realize that if you execute that connect() statement twice in your code, somewhere, your slot would get called twice? Are you saying that the sender sends different values each time and you see the same bytes twice as pairs (like 1,1,10,10,100,100,...)? Ah, no, it's 1 every time: suggest you change that (e.g. just increment each time) so you see different values. Does your dataRx() read whatever has been received? What is bytesAvailable() reporting each time? I am starting by trying to discover whether you are seeing twice as many bytes actually sent from Arduino vs readyRead() signal being fired twice vs your slot being called twice.

[Pl45m4]

@herrgross

I still assume the signals are coming in delayed (due to serialPort or signal event queue) and because of that your slot restarts the timer and therefore you have significantly shorter timespans in between.
Just a guess without analysing your code.

[herrgross]

@JonB I allready searched if I connected twice... could not find that.
here's some of the list I get:

and here's the code of my dataRx:

void Sirene::dataRx()
{
    if (port->bytesAvailable() > 0)
    {
        QByteArray data = port->readAll();
        QString value;
        value = QString::number(data[0]);
        statusBar()->showMessage(value);
        data.clear();
    }

    now = std::chrono::system_clock::now();

    rxTimepoints->push_back(now.time_since_epoch().count());
}

[JonB]

@herrgross
That's a lot of code; if you want to debug I would suggest a much smaller example.

You haven't shown whether the port->bytesAvailable() > 0 is followed every time dataRx() is called. You haven't shown the value you receive which I thought you were changing from always being 1 --- you send it to the statusbar, which isn't very helpful.

I don't understand your timing numbers (now, now-before), I don't see how that helps you tell whether readyRead() is being emitted twice per byte input.

I suggest you reduce this to simple code illustrating what is going on.

[herrgross]

@JonB ok, I reduced the code to what is necessary... (edited above)
the now and now - before in the list are :
now = std::chrono::system_clock::now();
and that value minus the previous now-value (which is the timespan between these two time_points)
So when the motor speeds up these values should continuously decrease, as the timespan between each call shoud be getting smaller.
I changed now the arduino, so it sends each time an increased value,
and in my statusbar (althoug not very pretty) I can see, that the value of QByteArray[0] is raising by 2 every call (so I get 2,4,6,8,10...)
that should not happen I guess...

[herrgross]

@Pl45m4 there is no timer being started by the slot...

[JonB]

@herrgross said in QSerialport readyRead() for measuring revolutions:
Originally you said:

Looking at this list I find that dataRx seems to be called twice per byte sent by the ARDUINO

Now you say:

I can see, that the value of QByteArray[0] is raising by 2 every call (so I get 2,4,6,8,10...)

which sounds more like "dataRx seems to be called only once per 2 bytes sent by the ARDUINO".

As I've said, you need to find out what the actual case is. Get that sorted out before you concentrate on your timings.

[herrgross]

@JonB what would you suggest how I can find out? I'm stuck in this since yesterday and run out of ideas...

[JonB]

@herrgross
I don't know what your problem is. I have not seen convincing evidence that readyRead() is being emitted any more (or less) than it should be. You could write simpler code to verify that if you wish, or not. Only you know what your timing figures mean, I do not. I don't know that readyRead() signal is a reliable means of measuring timings in real time. I don't understand hardware and have no idea what an Arduino is :)

[herrgross]

@JonB well, thanks for your help so far... I'll keep trying

[herrgross]

@herrgross got it!!!!
in the arduino code the interrupt must be set to RISING instead of HIGH, so it's only sending when the state of the interruptpin c h a n g e s to HIGH

  attachInterrupt(digitalPinToInterrupt(interruptPin), impuls, RISING);

thank you all!!